This message is in MIME format. Since your mail reader does not understand
this format, some or all of this message may not be legible.
------_=_NextPart_001_01C23971.F7684A50
Content-Type: text/plain;
charset="iso-8859-1"
If your drives are waiting on data to arrive you might not find much help in
messing with the buffer settings. Here's a technote, though, on the buffers
and how to use them. I found benefit in matching my SIZE_DATA_BUFFERS equal
to my NET_BUFFER_SZ.
http://seer.support.veritas.com/docs/183702.htm
<http://seer.support.veritas.com/docs/183702.htm>
One way to combat the drive waiting is to find a way to increase the data
arriving at the media server. Multiplexing backups, sending multiple client
streams at once, is one way to do this. I also have my backup server
connected with a Gig-E port, making it possible to collect data from more
than one standard lan 100 Mb lan client without bottlenecking at the
receiver.
Note, that if your backup server connection is at 100 Mbits/sec that's more
like 10 MBytes/sec. That's not much faster than a standard 7-8 Mbytes/sec
for a DLT drive.
-M
-----Original Message-----
From: Cierkowski, Wayne [mailto:Wayne.Cierkowski AT harpercollins DOT com]
Sent: Thursday, August 01, 2002 6:56 AM
To: 'veritas-bu AT mailman.eng.auburn DOT edu'
Subject: [Veritas-bu] Performance/Buffers
I've read about tuning the buffers in NB3.4.1 and have been attempting to
reduce the number of waits concerning empty buffers buy increasing the
number of data buffers file. But I still see a considerable number of waits
for full buffers. The way I understand this is that the tape drive is now
the one waiting and can greatly impact performance. But to reduce this wait
one needs to adjust the size of the data buffer. I run at 65536 which is
suppose to be the default for DLT 4000-8000 drives. I was told at a training
class to reduce the size of the data buffers to reduce the number of times
you wait for a full buffer. But in the training material it also states that
one should not go below the default size for the drive.
Do I accept what I have and just live with it or decrease the size of the
data buffer down to 32K versus 64K.
Wayne D. Cierkowski
Project Engineering Advisor
HarperCollinsPublishers
1000 Keystone Industrial Park
Scranton, PA 18512-4621
* Phone: (570) 941-1387
* Fax: (570) 941-1593
* e-Mail: Mailto:wayne.cierkowski AT harpercollins DOT com
<Mailto:wayne.cierkowski AT harpercollins DOT com>
***************************************************************
This message is intended only for the use of the individuals to which it is
addressed
and may contain information that is privileged and confidential. If you are
not the intended
recipient, you are hereby notified that you have received this transmission
in error;
any review, dissemination, distribution or copying of this transmission is
prohibited.
If you have received this communication in error, please notify us
immediately by reply
e-mail and delete this message and all of its attachments.
------_=_NextPart_001_01C23971.F7684A50
Content-Type: text/html;
charset="iso-8859-1"
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-1">
<TITLE>Performance/Buffers</TITLE>
<META content="MSHTML 6.00.2716.2200" name=GENERATOR></HEAD>
<BODY>
<DIV><SPAN class=525553015-01082002><FONT face=Arial color=#0000ff size=2>If
your drives are waiting on data to arrive you might not find much help in
messing with the buffer settings. Here's a technote, though, on the
buffers and how to use them. I found benefit in matching my
SIZE_DATA_BUFFERS equal to my NET_BUFFER_SZ.</FONT></SPAN></DIV>
<DIV><SPAN class=525553015-01082002><FONT face=Arial color=#0000ff
size=2></FONT></SPAN> </DIV>
<DIV><SPAN class=525553015-01082002><FONT face=Arial color=#0000ff size=2><A
href="http://seer.support.veritas.com/docs/183702.htm">http://seer.support.veritas.com/docs/183702.htm</A></FONT></SPAN></DIV>
<DIV><SPAN class=525553015-01082002><FONT face=Arial color=#0000ff
size=2></FONT></SPAN> </DIV>
<DIV><SPAN class=525553015-01082002><FONT face=Arial color=#0000ff size=2>One
way to combat the drive waiting is to find a way to increase the data arriving
at the media server. Multiplexing backups, sending multiple client
streams
at once, is one way to do this. I also have my backup server connected
with a Gig-E port, making it possible to collect data from more than one
standard lan 100 Mb lan client without bottlenecking at the
receiver.</FONT></SPAN></DIV>
<DIV><SPAN class=525553015-01082002><FONT face=Arial color=#0000ff
size=2></FONT></SPAN> </DIV>
<DIV><SPAN class=525553015-01082002><FONT face=Arial color=#0000ff size=2>Note,
that if your backup server connection is at 100 Mbits/sec that's more like
10 MBytes/sec. That's not much faster than a standard 7-8 Mbytes/sec for
a
DLT drive.</FONT></SPAN></DIV>
<DIV><SPAN class=525553015-01082002><FONT face=Arial color=#0000ff
size=2></FONT></SPAN> </DIV>
<DIV><SPAN class=525553015-01082002><FONT face=Arial color=#0000ff
size=2>-M</FONT></SPAN></DIV>
<BLOCKQUOTE>
<DIV class=OutlookMessageHeader dir=ltr align=left><FONT face=Tahoma
size=2>-----Original Message-----<BR><B>From:</B> Cierkowski, Wayne
[mailto:Wayne.Cierkowski AT harpercollins DOT com]<BR><B>Sent:</B> Thursday,
August
01, 2002 6:56 AM<BR><B>To:</B>
'veritas-bu AT mailman.eng.auburn DOT edu'<BR><B>Subject:</B> [Veritas-bu]
Performance/Buffers<BR><BR></FONT></DIV>
<P><FONT face="Times New Roman">I've read about tuning the buffers in NB3.4.1
and have been attempting to reduce the number of waits concerning empty
buffers buy increasing the number of data buffers file. But I still see a
considerable number of waits for full buffers. The way I understand this is
that the tape drive is now the one waiting and can greatly impact
performance.
But to reduce this wait one needs to adjust the size of the data buffer. I
run
at 65536 which is suppose to be the default for DLT 4000-8000 drives. I was
told at a training class to reduce the size of the data buffers to reduce the
number of times you wait for a full buffer. But in the training material it
also states that one should not go below the default size for the drive.
</FONT></P>
<P><FONT face="Times New Roman">Do I accept what I have and just live with it
or decrease the size of the data buffer down to 32K versus 64K.</FONT> </P>
<P><B><FONT face="Comic Sans MS">Wayne D. Cierkowski</FONT></B> <BR><FONT
face="Times New Roman">Project Engineering Advisor</FONT> </P>
<P><FONT face=Garamond color=#ff0000>Harper</FONT><FONT face=Garamond
color=#0000ff>Collins</FONT><I></I><I><FONT face=Stempel
color=#808080>Publishers</FONT></I> <BR><FONT face=Tahoma size=1>1000
Keystone
Industrial Park</FONT> <BR><FONT face=Tahoma size=1>Scranton, PA
18512-4621</FONT> </P>
<P><FONT face=Wingdings size=2>(<FONT face="Courier New"></FONT></FONT> <FONT
face=Arial size=2>Phone: (570) 941-1387</FONT> <BR><FONT face=Wingdings
size=2>(<FONT face="Courier New"></FONT></FONT> <FONT face=Arial size=2>Fax:
(570) 941-1593</FONT> <BR><FONT face=Wingdings size=2>*<FONT
face="Courier New"></FONT></FONT> <FONT face=Arial
size=2>e-Mail:<U></U></FONT><U></U><U><FONT face="Times New Roman"
size=2></FONT> <FONT face=Arial color=#0000ff size=2><A
href="Mailto:wayne.cierkowski AT harpercollins DOT
com">Mailto:wayne.cierkowski AT harpercollins DOT
com</A></FONT></U><BR></P><BR><CODE><FONT
size=3><BR><BR>***************************************************************<BR>This
message is intended only for the use of the individuals to which it is
addressed <BR>and may contain information that is privileged and
confidential.
If you are not the intended <BR>recipient, you are hereby notified that you
have received this transmission in error; <BR>any review, dissemination,
distribution or copying of this transmission is prohibited.<BR>If you have
received this communication in error, please notify us immediately by reply
<BR>e-mail and delete this message and all of its
attachments.<BR></BLOCKQUOTE></FONT></CODE></BODY></HTML>
------_=_NextPart_001_01C23971.F7684A50--
|